Phishing Predictor

September 30, 2016
Uses Two-class NN, Decision Jungle and Boosted trees to predict if a site is a phishing site or not. Dataset from UCI
I was browsing twitter and saw a tweet from Nicholas Papernot on how he created simple tutorial to classify phishing sites using the UCI dataset (It is awesome, and you should check it out. Link below) I recreated his tutorial in Azure ML. I wanted fast and accurate learners, so I used Boosted trees, Decision Jungles and a simple two class Neural Network. Accuracy Results: ----------------- - Boosted Trees: 96.6% - Decision Jungles: 94.3% - Two class Neural Network: 95.9% Reference --------- 1. Original Tutorial idea - @NicholasPapernot - 2. @NicholasPapernot awesome tutorial - [][1] 3. Dataset Available at [1]: